Stress Analysis of Pump Piping (Centrifugal) System using Caesar II

Every process piping industry uses several pumps in each process unit. Sometimes the analysis is very critical. In this article I will try to elaborate the method followed for stress analysis of a centrifugal pump piping system.  The stress system consists of typical discharge lines of two centrifugal pump (Pump A and Pump B). Fluid from this two pumps are pumped into a heat exchanger. As per P&ID only one pump will operate at a time, other pump will be a stand by pump. I will explain the stress analysis methodology in three parts:- a) Modeling of Pump b) Preparation of analysis Load cases  and c) Analysing the output results.

pump piping

                                           Fig. 1: Sample pump piping model as it looks in Caesar II

A. Modeling of Pump:

For modeling the pump we require vendor general arrangement drawing or outline drawing. All rotary equipments are modeled as a weightless rigid body in Caesar II. From the outline drawing we need to take the dimensions till some fixed point. Let us take the example of the outline drawing shown in figure 2.

Pump Outline Drawing

                                    Fig. 2: Sample outline drawing for a centrifugal pump

From the above drawing we can get the dimensions for elements 10-5000 as 8.5 inch and element 5000-5020 as 6.19 inch. At node 5020 we will provide fixed anchor. During modeling of the above elements we need to use line size and thickness as diameter and thickness of the equipment. Line temperature and pressures as equipment properties. We have to provide anchor (with cnode) at node 10 for checking nozzle loads which we will compare with the allowable value as provided in Fig. 3 below:

allowable load

                      Fig. 3: Allowable nozzle load values as mentioned in Equipment GA drawing

In absence of allowable load value the Pump design code (API 610 for API pumps, ANSI HI 9.6.2 for non API pumps) can be followed for the same.

After the pump is modeled as rigid body the piping modeling need to be done from pump-piping interconnection flanges.

B. Preparation of Analysis Load Cases:

Along with normal load cases two additional load cases need to be prepared. Normally in refinery and petrochemical industry one pump operates and other acts as a stand by pump. So we have prepare load cases as follows:

1. Hydrostatic case (WW+HP                    HYD)

2. Operating case with both pump operating (W+T1+P1                  OPE)

3. Operating case with total system in maximum design temperature ( W+T2+P1                  OPE)

4. Operating case with pump A operating and pump B Stand by (W+T3+P1           OPE)

 5. Operating case with pump B operating and pump A Stand by (W+T4+P1           OPE)

6. Operating case with total system in minimum design temperature ( W+T5+P1                  OPE)

Next all normal load cases like static seismic, static wind, etc are to be built as per stress analysis or flexibility specification.

When pump A is in operating condition and pump B stand by then normal pipe operating temperature has to be inserted till Tee connection for pump A and ambient temperature will be the input input for pump B as shown in Fig. 4.  Similarly reverse the input when pump B is operating.

temperature profile

                      Fig. 4: Operating-Stand By Temperature profile for two pump system

After equipment is modeled completely start  modeling the piping following dimensions from piping isometric drawings. Try to make a closed system. Normally pump lines are connected to some vessel, tank or heat exchangers. So it will create a close system. Then run the analysis to check stresses, displacements and loads.

C. Analysing the output Result:

Once Caesar completes its iteration process we can see the output results in output window. At nozzles (the nodes which we anchored with a cnode) we can check the force values. These values we have to compare with the allowable values. If the actual values are less than the allowable values then the nozzle is safe. Otherwise we have the make changes in supporting or routing to bring the nozzle load values within allowables. A sample output restraint is provided in Fig. 5 for your reference.

Restraint Summary

                                              Fig 5. Typical output results for two pump system

As can be seen from the above figure that we are checking nozzle loads in load case 2, 4, 5 and 6. For rotary equipments normally nozzle qualification in design or upset temperature is not required.

Special Consideration for Rotary Equipments:

Now we have to make one separate caesar file and we have to check sustained displacement at nozzle at WNC (weight no content) case. This checking will ensure proper alignment of piping flange and equipment nozzle flange. For detailed analysis steps follow this link:

Always remember to provide first piping support from pump nozzle as an adjustable support (or a spring support) to aid in alignment.

In case of 3 pump system, normally two pumps will be operating and one pump will be stand by. So input and prepare load cases accordingly.

If you have any confusion or want to add more please write in comments section.

Anup Kumar Dey

I am a Mechanical Engineer turned into a Piping Engineer. Currently, I work in a reputed MNC as a Senior Piping Stress Engineer. I am very much passionate about blogging and always tried to do unique things. This website is my first venture into the world of blogging with the aim of connecting with other piping engineers around the world.

18 thoughts on “Stress Analysis of Pump Piping (Centrifugal) System using Caesar II

  1. the temperature distribution between Isolating valve and Tee section ambient of stagnant condition(I consider (T+Tamp)/2)?
    when Pump B is stand by it doesn’t make any effect? while Pump A is standby itz make influence on nozzle as it is long section.

    1. Average temp philosophy is considered if the line is insulated. Otherwise in the long run the temp will drop and reach almost ambient condition..

  2. I’ve learn a few good stuff here. Certainly worth bookmarking for revisiting.
    I wonder how a lot effort you put to make
    one of these excellent informative website.

  3. It’s not my first time to pay a visit this website,
    i am browsing this website dailly and get good information from here
    every day.

  4. hello there and thank you for your info – I’ve certainly picked up
    anything new from right here. I did however expertise several technical
    points using this website, since I experienced to reload the web site lots of
    times previous to I could get it to load properly.
    I had been wondering if your hosting is OK? Not that I am complaining, but slow loading instances times will often affect your placement in google and
    could damage your quality score if advertising and
    marketing with Adwords. Well I am adding this RSS to my email and can look out for much more of your
    respective exciting content. Make sure you update this again soon.

  5. I can not understand the concept behind considering temp. T3, T4 T5.
    Usually we consider same temp. for both the pumps , will it make any difference ?

  6. Few Nice insights on this topic is provided by Mr. David Diehl and Mr. Edward Klein in linkedin. The original comments by them can be accessed by following the link mentioned at the end. However for convenience of the readers I am producing those in this section. 

    By Mr. Edward Klein:
    Senior Pipe Stress Engineer at S & B Engineers and Constructors, Ltd.

    While the overall article is good, I do have a couple of points. 
    First, the guides shown on the base supports in front of the pump are a modeling practice that I strongly discourage. These guides are a “paper solution” as their proximity to the pumps, along with a typical installation gap and relatively low stiffness compared to the pump casing means that any loads that Caesar shows them carrying will in reality be applied to the pump nozzles. We can certainly make a case that pump vendors excessively handicap us with the allowables published in API and ANSI, but they are what we have to work with. Those guides may show you qualifying in your model, while the actual pumps are over loaded. 
    Also, while we also employ as standard the modeling practice of treating the piping for the idle pump as ambient, there are cases with hotter pump systems that need to be considered. Often, hot operating systems will have a warm up bypass around the check valve to backflow a small stream through the idle pump. The purpose of this is to keep the idle pump in a ready condition and to avoid the potential for cracking induced by thermal shock. In these systems, we take the normal case for nozzle qualification with both pumps at temperature. There can still be times when one side is closer to ambient if the pump is taken out for maintenance, but these become more of a temporary upset case, rather than the long term exposure to loads that the allowables are intended to protect the pumps from. 
    Finally, I’ve generally moved away from the habit of modeling a rigid vertical element from the pump center to the base. API horizontal pumps are centerline supported anyway. ANSI pumps are foot mounted. Yet, even in this case, I find it creates an exaggerated effect due to the stiffness values of restraints and rigids in the model and the fact that we typically model piping as supported at its centerline, even though the support points are at the bottom of the pipe.

    By Mr. David Diehl
    Director of Training at Intergraph|Training & Communications Executive|Leadership & Organizational Development

    CAESAR II now provides a nozzle check report in the static output that compares identified nozzle loads to limits provided in standards. This screening report is more convenient than referencing the restraint report as you show. 
    Also, some would question the sudden temperature changes defined for the header but your approach is conservative and similar to the recommendations in WRC 449.

    The actual comments can be accessed from the below mentioned link:

  7. Normally  for Nozzle load checking shall we consider the seismic effect and wind effects loads case also…
    give me the suggestion pls

  8. I do not see many comments here, it means you have low visitors. I know how to make your page go viral. If you want to know just search in google for:
    Kimting’s Method To Go Viral

  9. Problem:
    Of a 3-pump systems, 2-pumps are runnng at all time while the third one is a standby.
    All three pumps are shaking violently like they will be thrown of off the foundations.

    Caesar stress analysis recommends some piping and support modifications based on the results.
    My question is how wise it is to delve into modifications based on the result recommendations?

    Shouldn’t we do CFD as well and then should perform some kind of correlation study before we spend any money on the actual modifications?

  10. as I heard from my senior, it is common practice to multiplied by 4 of line size and thickness too in modelling rigid pump. how about your comment for this statement? thanks sir.

Leave a Reply

Your email address will not be published. Required fields are marked *


Recent Content

Subscribe For Latest Updates

Signup for our newsletter and get notified when we publish new articles for free!